The general technologies of targeting mutations into the genome of cells, and the process of generating mouse lines from genetically altered embryonic stem (ES) cells with specific genetic lesions are well known (Bradley, 1991, Cur. Opin. Biotech. 2:823-829). A random method of generating genetic lesions in cells (called gene, or promoter, trapping) has been developed in parallel with the targeted methods of genetic mutation (Allen et al., 1988 Nature 333(6176):852-855; Brenner et al., 1989, Proc. Natl. Acad. Sci. U.S.A. 86(14):5517-5521; Chang et al., 1993, Virology 193(2):737-747; Friedrich and Soriano, 1993, Insertional mutagenesis by retroviruses and promoter traps in embryonic stem cells, p. 681-701. In Methods Enzymol., vol. 225., P. M. Wassarman and M. L. DePamphilis (ed.), Academic Press, Inc., San Diego; Friedrich and Soriano, 1991, Genes Dev. 5(9):1513-1523; Gossler et al., 1989, Science 244(4903):463-465; Kerr et al., 1989, Cold Spring Harb. Symp. Quant. Biol. 2:767-776; Reddy et al., 1991, J Virol. 65(3):1507-1515; Reddy et al., 1992, Proc. Natl. Acad. Sci. U.S.A. 89(15):6721-6725; Skarnes et al., 1992, Genes Dev. 6(6):903-918; von Melchner and Ruley, 1989, J. Virol. 63(8):3227-3233; Yoshida et al., 1995, Transgen. Res. 4:277-287). Gene trapping provides a means to create a collection of random mutations by inserting fragments of DNA into transcribed genes. Insertions into transcribed genes are selected over the background of total insertions since the mutagenic DNA encodes an antibiotic resistance gene or some other selectable marker. The selectable marker lacks its own promoter and enhancer and must be expressed by the endogenous sequences that flank the marker after it has integrated. Using this approach, transcription of the selectable marker is activated and the cell gene is concurrently mutated. This type of strict selection makes it possible to easily isolate thousands of ES cell colonies, each with a unique mutagenic insertion.
Collecting mutants on a large-scale has been a powerful genetic technique commonly used for organisms which are more amenable to such analysis than mammals. These organisms, such as Drosophila melanogastor, yeast Saccharomyces cerevisiae, and plants such as Arabadopsis thalia are small, have short generation times and small genomes (Bellen et al., 1989, Genes Dev. 3(9):1288-1300; Bier et al., 1989, Genes Dev. 3(9):1273-1287; Hope, 1991, Develop. 113(2):399-408. These features allow an investigator to rear many thousands or millions of different mutant strains without requiring unmanageable resources. However, these type of organisms have only limited value in the study of biology relevant to human physiology and health. It is therefore important to have the power of large-scale genetic analysis available for the study of a mammalian species that can aid in the study of human disease. Given that the entire human genome is presently being sequenced, the comprehensive genetic analysis of a related mammalian species will provide a means to determine the function of genes cloned from the human genome. At present, rodents, and particularly mice, provide the best model for genetic manipulation and analysis of mammalian physiology.
Gene trapping has been used as an analytical tool to identify genes and regulatory regions in a variety of animal cell types. One system that has proved particularly useful is based on the use of ROSA (reverse orientation splice acceptor) retroviral vectors (Friedrich and Soriano, 1991 and 1993).
The ROSA system can generate mutations that result in a detectable homozygous phenotype with a high frequency. About 50% of all the insertions caused embryonic lethality. The specifically mutated genes may easily be cloned since the gene trapping event produces a fusion transcript. This fusion transcript has trapped exon sequences appended to the sequences of the selectable marker allowing the latter to be used as a tag in polymerase chain reaction (PCR)-based protocols, or by simple cDNA cloning. Examples of genes isolated by these methods include a transcription factor related to human TEF-1 (transcription enhancer factor-1) which is required in the development of the heart (Chen et al., 1994, Genes Devel. 8:2293-2301. Another (spock), is distantly related to yeast genes encoding secretion proteins and is important during gastrulation.
The above experiments have established that the ROSA system is an effective analytical tool for genetic analysis in mammals. However, the structure of many ROSA vectors selects for the "trapping" of 5' exons which, in many cases, do not encode proteins. Such a result is adequate where one wishes to identify and eventually clone control (i.e., promoter or enhancer) sequences, but is not optimal where the generation of insertion-inactivated null mutations is desired, and relevant coding sequence is needed. Thus, the construction of large-scale mutant (preferably null mutant) libraries requires the use of vectors that have been designed to select for insertion events that have occurred within the coding region of the mutated genes as well as vectors that are not limited to detecting insertions into expressed genes.